`nimare.utils`.unique_rows

unique_rows(ar, return_counts=False)[source]

Remove repeated rows from a 2D array.

In particular, if given an array of coordinates of shape (Npoints, Ndim), it will remove repeated points.

Parameters:

ar (2-D ndarray) – The input array.
return_counts (bool, optional) – If True, also return the number of times each unique item appears in ar.

Returns:

ar_out (2-D ndarray) – A copy of the input array with repeated rows removed.
unique_counts (np.ndarray, optional) – The number of times each of the unique values comes up in the original array. Only provided if return_counts is True.

:raises ValueError : if ar is not two-dimensional.:

Notes

The function will generate a copy of ar if it is not C-contiguous, which will negatively affect performance for large input arrays.

This is taken from skimage. See skimage.util.unique_rows().

Examples

>>> ar = np.array([[1, 0, 1],
...                [0, 1, 0],
...                [1, 0, 1]], np.uint8)
>>> unique_rows(ar)
array([[0, 1, 0],
    [1, 0, 1]], dtype=uint8)

License

nimare.utils.unique_rows

`nimare.utils`.unique_rows